A nonparametric copula approach to predict heart rate trajectories from single overnight wearable accelerometry

Area of Science:

• Biomedical Engineering
• Health Informatics
• Wearable Technology

Background:

• Wearable devices enable continuous, non-invasive monitoring of physiological data like heart rate and motion.
• Current heart rate prediction methods often rely on proprietary algorithms or specific physiological signals.
• There is a need for advanced, accessible methods for heart rate monitoring in precision health.

Purpose of the Study:

• To develop and evaluate a novel method for estimating heart rate trajectories using only overnight triaxial accelerometer data.
• To assess the performance of a copula-based transformation with a local log-quadratic likelihood kernel estimator and nearest-neighbor bandwidth (Tll2nn).
• To demonstrate the feasibility of heart rate tracking without direct physiological signal input.

Main Methods:

• Utilized a copula-based transformation method combined with a local log-quadratic likelihood kernel estimator and nearest-neighbor bandwidth (Tll2nn).
• Applied the method to overnight triaxial accelerometer data from Apple Watch devices.
• Tested the model on data from approximately 30 healthy participants during sleep.

Main Results:

• The Tll2nn model effectively captured nonlinear dependencies between motion-derived features and heart rate fluctuations.
• The proposed method outperformed a baseline prediction of constant mean heart rate in 20 out of 26 subjects.
• Demonstrated accurate heart rate trajectory tracking using only overnight motion data.

Conclusions:

• The developed methodology provides a novel, interpretable, and computationally efficient approach for automated heart rate tracking.
• This technique enables scalable automation of heart rate monitoring using readily available wearable device data.
• The findings support the potential for revolutionizing remote patient monitoring, early detection of irregularities, and management of sleep and cardiovascular conditions.